一种同质传感器的最优加权融合算法研究

Research on Optimal Weight Fusion Algorithm of Homogeneous Sensors

为了满足同质传感器采集数据的可靠性需求,提升算法的效率,提出了一种改进的最优加权融合算法。基于3sigma准则,使用估计标准差代替实际标准差剔除粗大误差值,从而提高数据融合的可靠性。利用该算法对一氧化碳的原始数据进行了预处理,结果表明与格拉布斯准则、狄克逊准则等相关算法相比,新算法对粗大误差的剔除更准确,效率更高,减小了待融合数据的标准差。

In order to meet the reliability requirement of homogeneous sensor data acquisition and improve the efficiency of the algorithm,the optimal weighted fusion algorithm was proposed.Based on the 3sigma criterion,the estimated standard deviation was used to replace the actual standard deviation to eliminate the outliers to improve the reliability of data fusion.The algorithm was used to preprocess the raw data of carbon monoxide.The results show that compared with the Grubbs criterion and Dixon criterion,the new algorithm is more accurate and efficient in eliminating the outliers and the standard deviation of the fused data can be decreased.